Migrate tests/test_core.py to npt.assert_allclose

[viknesh-ai]

Related to #1175

Continuing the file-by-file split — this one covers tests/test_core.py. Largest and messiest file in the split so far: 131 call sites, and unlike the other files, the naming here was inconsistent throughout (some pairs ref/comp, some ref/cmp, some ref_X/comp_X, and a few reversed as suffix forms like P_ref/P_comp, I_ref/I_comp).

assert_almost_equal only checks a fixed absolute tolerance and NumPy's docs recommend assert_allclose instead. What changed:

• Normalized every pair to the ref_X/cmp_X prefix convention, including the reversed suffix forms (P_ref/P_comp -> ref_P/cmp_P, I_ref/I_comp -> ref_I/cmp_I, mp_ref -> ref_mp, etc.) and a couple of comp-not-cmp stragglers. Did a token-level rename (not a blind substring replace) to avoid clobbering unrelated words like compare/compute/comparison in comments.
• All 131 original calls were already in (ref, comp) order, so every one needed flipping to (cmp_side, ref_side).
• None of the 131 sites used decimal= or referenced config.STUMPY_TEST_PRECISION — all used numpy's implicit decimal=7 default, so all 131 use the literal atol=1.5e-07.
• Checked the actual runtime dtype of every distinct naive.* source feeding a ref_ side before touching anything. Only naive.find_matches/core._find_matches (2 call sites, in test_find_matches_all/test_find_matches_maxmatch) come back dtype=object (a [distance, index] pair array, same root cause as the earlier files) — cast both sides to float64 there. Everywhere else the naive source (compute_mean_std, rolling_nanstd, distance_matrix, rolling_isconstant, get_array_ranges, rolling_window_dot_product, merge_topk_PI/merge_topk_ρI, and this file's own local naive_idx_to_mp/naive_bfs_indices helpers) is already clean float64/int64, so no unnecessary casting was added.
• Also applied the same ref_/cmp_ naming fix to a handful of npt.assert_array_equal calls that shared the same inconsistent naming (test_apply_exclusion_zone*, test_get_mask_slices, test_gpu_searchsorted, test_process_isconstant_*), per the standing rule that this convention applies to assert_array_equal too, not just assert_allclose.
• One non-naive.* pair (test_array_to_temp_file, a value round-tripped through core.array_to_temp_file + np.load) got the same treatment: the original value is ref_val, the round-tripped value is cmp_val. Left the unrelated left/right pairs in test_jagged_list_to_array*/test_replace_distance untouched since those aren't naive-vs-stumpy comparisons and are out of scope for this migration.
• A few commented-out # npt.assert_almost_equal(...) lines got their syntax updated too for consistency, staying commented.

Ran the full file twice locally (normal mode and JIT-disabled/CUDA-simulated, matching test.sh coverage) — 244 passed + 1 skipped (GPU test, no CUDA available locally) in normal mode, 245 passed in CI-equivalent mode (the GPU test runs under the CUDA simulator there).

To Do List

Standing checklist from @seanlaw for this migration, applies to every file in the split - status below is for test_core.py:

• Each test file correctly replaces all uses of npt.assert_almost_equal with its equivalent npt.assert_allclose
• When rtol=0 for npt.assert_allclose, simply omit this parameter and value since this is the default
• Ensure that npt.assert_allclose(actual, desired) always has the stumpy computed value as actual and the naive computation as desired
• Use atol=1.5e-07 rather than atol=1.5*10**-07 - applies to all 131 sites here (none referenced config.STUMPY_TEST_PRECISION)
• Eventually, replace the ugly 1.5*10**-config.STUMPY_TEST_PRECISION with config.STUMPY_TEST_PRECISION = 1.5e-07 (in config.py) - tracked as a follow-up, not part of this per-file migration (n/a for this specific file, but keeping the item open since it's still pending overall)
• Use cmp instead of comp - this file had several comp/comp_X stragglers, all renamed
• Rename naive outputs to ref_ and stumpy-computed outputs to cmp_ (applies to npt.assert_array_equal too) - this file had the most inconsistent naming of the series (including reversed suffix forms like P_ref/P_comp), fully normalized

Pull Request Checklist

Below is a simple checklist but please do not hesitate to ask for assistance!

• Read our Contributing Guide
• Referenced a Github issue (or create one if one doesn't already exist)
• Read and reviewed all of the comments in the Github issue that you've referenced (along with cross-referenced issues/pull requests) to ensure that the issue still requires a pull request
• Checked that the issue has not already been assigned to anybody else or is already being addressed in another pull request
• Left a meaningful comment on the original Github issue to discuss the detailed approach for your contribution and received confirmation from the maintainers before proceeding with this pull request
• Forked, cloned, and checked out the newest version of the code
• Created a new branch
• Made necessary code changes
• Installed black (i.e., python -m pip install black or conda install -c conda-forge black)
• Installed flake8 (i.e., python -m pip install flake8 or conda install -c conda-forge flake8)
• Installed pytest-cov (i.e., python -m pip install pytest-cov or conda install -c conda-forge pytest-cov)
• Ran black --exclude=".*\.ipynb" --extend-exclude=".venv" --diff ./ in the root stumpy directory
• Ran flake8 --extend-exclude=.venv ./ in the root stumpy directory
• Ran ./setup.sh dev && ./test.sh in the root stumpy directory and ensured that all tests are passing locally
• Check this box if AI code assistance was used to generate 15%+ of the code in this pull request

Please do not commit any code to avoid/circumvent a failing test and, instead, engage in a discussion (below) to determine the best course of action.

Only request a review after the checklist above is fully completed!

[gitnotebooks]

Review these changes at https://app.gitnotebooks.com/stumpy-dev/stumpy/pull/1188

[viknesh-ai]

@seanlaw next file in the split — same conventions as the previous PRs. Ready for review whenever you have time.

[seanlaw]

For assert_array_equal, please also ensure that switch cmp is first and ref is second

[seanlaw]

npt.assert_allclose and switch ref/cmp order

[seanlaw]

switch ref/cmp order

[seanlaw]

switch ref/cmp order

[seanlaw]

switch ref/cmp order

[seanlaw]

switch ref/cmp order

[seanlaw]

switch ref/cmp order

[seanlaw]

switch ref/cmp order